I am trying to merge the NHIS public used perone file with IHIS data for both 1991 and 1992 survey. However part of the data from NHIS 1992 has the year variable equals 1991. Therefore, when creating the unique idtifier NHISPID by the instruction in the NHIS data, these observations in 1992 survey, which has the year variable equals 1991, has duplicated NHISPID with the observation in 1991 survey. How should I deal with it.
Linking data between the IPUMS Health Surveys NHIS file with the public use file can be tricky in these pre-1998 files, however, IPUMS Health has detailed documentation for performing these links, which can be found in the “Linking” user note.
As you can see from the tables on that page, the 1992 file is a bit different from it’s surrounding years and is actually the only one of the constructed identifiers to use year.
I hope this helps.
I see. I actually need both “year” and “nhispid” to merge multi-years data. That sovles the problem. Thanks.